Arpit Dalal

435 posts

Arpit Dalal banner
Arpit Dalal

Arpit Dalal

@arpitdalal_dev

SWE @plusgrade | @iHariPrabodham devotee and volunteer

Canada Beigetreten Haziran 2018
80 Folgt41 Follower
Angehefteter Tweet
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
OMG!!! What?!? I’m on the Epic Stack contributors slide presented by @kentcdodds at the Epic Web Conf I’ve learned so much from the Epic Stack, and it’s been amazing to give back to such an incredible project. Huge thanks to Kent for all the awesome educational content!
Arpit Dalal tweet media
English
2
0
5
820
Arpit Dalal retweetet
Poke
Poke@interaction·
Starting today, personal superintelligence is just one tap away. No download, no signup. Text Poke for free now: Poke.com 🌴 — 0:00 – What's Poke? 0:50 – Introducing Poke Recipes 1:25 –  Create a Recipe in 10 seconds 1:43 – Earn on Poke 2:44 – Build with npx poke 12:58 – Recap 13:36 – Parisian Love
English
167
116
1.2K
503.9K
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
After the latest Codex update, it can’t run any scripts and complains about node/npm not being in the shell. I’am running 5.4 locally with full access. Restart didn’t fix it 😞 This has broken the feedback loop @OpenAIDevs! Agent can’t verify its own work and I’ve to babysit it constantly.
Arpit Dalal tweet media
English
0
0
0
36
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
@wesbos one man’s poison is another man’s medicine
English
0
0
0
498
Wes Bos
Wes Bos@wesbos·
Me: why would anyone watch sports, what a waste of time! Also me: ohhh what did the one hosting company employee say to the other hosting company employe today?
English
18
12
530
37K
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
5.4 is a trash model at UI. whoever says otherwise is a liar.
English
0
0
0
4
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
I haven’t checked this site all day today 🤯
English
0
0
0
4
Supreme Leader Wiggum
Supreme Leader Wiggum@ScriptedAlchemy·
Cursors debug agent is one of the best pieces of engineering I’ve seen.
English
13
6
136
14.5K
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
@aidenybai Oops… did an agent come up with this?
English
0
0
0
354
Aiden Bai
Aiden Bai@aidenybai·
we had this insane P0 bug users would get randomly logged out we had a bunch of users churn... ...turns out Next.js prefetching the /logout page
English
91
24
1.8K
605K
Theo - t3.gg
Theo - t3.gg@theo·
Had a big migration that Codex absolutely bombed. Couldn’t even make the code compile. Opus one shot it. Both good models. Both far from perfect. You should use both and push their limits.
English
182
32
2.1K
105K
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
@theo I know it says “100% remaining” but… I think you should flip it and say “0% used” just because the title is “Usage”. Human mind reads “Usage” and the number “100%” and that creates the confusion. If it were a generic title like “Tokens”, maybe it should’ve been fine.
English
0
0
32
2.5K
Theo - t3.gg
Theo - t3.gg@theo·
I'm 🤏 to crashing out I swear
Theo - t3.gg tweet media
Theo - t3.gg@theo

T3 Chat Launch Week officially starts on Monday, but we wanted to get this out first. We just overhauled how credits work on T3 Chat. No more "standard" and "premium" credits. Now you have a "usage" bar that resets every 4 hours. You'll never be more than a few hours away from sending more messages. This change is meant to align your usage with our actual costs. Using a fancy model for a simple question? Costs almost nothing. Using Gemini 3 Pro to analyze giant PDFs in 20-message threads? That will eat your usage fast. For reasonable users, this should be a SIGNIFICANT increase in how many messages you can send per month. The goal of this change was to balance out costs and reduce "fear of running out". If you're not dumping your entire codebase into T3 Chat as context, you should never have to worry about limits again. For those who previously purchased premium credits, they are still available and will be used when you go over your existing limits. We also have a small gift for those who used a lot of premium credits (check your email later today for more info) These changes unblock some very exciting things in the near future, so keep your eyes peeled for that 👀

English
101
6
1K
211.6K
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
Cursor needs to fix something with their terminals. Codex has a great skill for addressing pr comments, it works beautifully 🤌 Cursor sees the skill but can’t use it because it doesn’t see that I’m authenticated via gh cli for some reason. Is this a skill issue? I tried running it outside of the sandbox mode but that didn’t help.
English
0
0
0
33
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
I switched back to Cursor today because of its Plan mode, created thorough plans with Opus 4.6 and implemented using Auto. Got a lot of work done. Don’t think I would’ve gotten so much done with Codex.
English
0
0
0
43
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
My Following tab is totally broken 😫
English
0
0
0
54
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
Mind blowing that 5.3-codex in Codex harness migrated entire Electron app to Tauri in 1 shot. It ran for hours. Wrote tests, updated GH actions, updated docs, everything in 1 shot. App worked perfectly 🤌 Don’t think any other model can do what it did in a single shot
English
1
0
0
38
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
Mind blowing that what 5.3-codex high in Codex harness couldn’t do after multiple back and forth for over 5 hours, Cursor did it in 1 shot under 30 mins including planning with Opus 4.6 and implementing it using Auto 🤯
English
1
0
0
176
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
Why has no one copied @cursor_ai’s debugging mode? It’s just sooooooo good 🤌 I want @OpenAI Codex with Cursor’s debugging mode and browser integration like yesterday 😫
English
0
0
0
418
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
@housecor According to theo, 5.3 gives more output than 5.2. Taking that into account, on paper, 5.2 would be better for throw-away code since little output means low context use and low output token price charged.
English
1
0
2
276
Cory House
Cory House@housecor·
tldr: If you haven’t used Codex lately, try again with the new 5.3 model. It responds faster, one shots more, and often provides better results than Opus. 5.3 is the first time I’m considering switching from Claude Code to Codex as my primary.
Theo - t3.gg@theo

I've been using 5.3 Codex for 3 weeks. It's an incredible model. I've built so much stuff with it. Made a vid showing everything I love about it, as well as a few call-outs of things I hope OpenAI changes.

English
43
15
269
29.2K
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
And it was removing duplicates from a pbxproj file, that’s all!
English
0
0
0
13
Arpit Dalal
Arpit Dalal@arpitdalal_dev·
Opus 4.6 got hung on a task that GPT 5.2-codex finished in one go. Notice I didn’t say GPT 5.3-codex because it’s not available over API yet.
English
1
0
0
73

Entdecken

@OpenAIDevs @wesbos @ScriptedAlchemy @aidenybai @theo @elonmusk @BarackObama @taylorswift13